From: Noveck, Dave (dave.noveck@netapp.com)
Date: 07/29/99-11:50:28 AM Z
Message-ID: <7F608EC0BDE6D111B53A00805FA7F7DA05D10B84@tahoe.corp.netapp.com> From: "Noveck, Dave" <dave.noveck@netapp.com> Subject: RE: Draft minutes for Oslo NFSv4 WG meeting. Date: Thu, 29 Jul 1999 09:50:28 -0700 > > > > The following are draft minutes for the NFSv4 meeting. > > > Please let me know if any corrections need to be made. > > > > > > Spencer > > > > > > --- > > > DRAFT Minutes, NFS version 4 (nfsv4) WG, Oslo IETF, > reported by Brent > > > Callaghan (edited by Spencer Shepler). > > > > ... > > > > re ACL's, > > > > > Based on the options presented there was not a clear > consensus on what > > > direction to undertake. While ACL support seemed to be strong > > > requirement for NFSv4, a clear solution or direction has > not presented > > > itself. > > > > It depends on how you define "clear", but I thought there was a > > clear consensus on the direction to take, although not a clear > > solution. At least I felt encouraged on that score. Of course, > > I felt even more encouraged when Carl Beame volunteered to take > > this on. Thanks, Carl. > > > > I would describe the consensus for the direction as follows: to > > take the NT model as fundamental, to take away things from of it > > of only marginal utility (e.g. many of the permissions bits that > > aren't normally separately set) while adding to the small set > > of things that were in the UNIX ACL models but not in NT model. > > The sense of the group was that in the event there was a conflict > > (i.e. some UNIX thing that just couldn't be fit in the NT model), > > the UNIX/POSIX feature should be considered dispensable. > > While this last statement represents my personal view, when I > asked the > attendees if they felt we should be willing to sacrifice UNIX > ACLs when such a > conflict exists, the reaction was cool at best. (It could be > that it didn't > phrase the question very well). My impression was there that were was > consensus at that meeting that existing UNIX ACL models need > to be preserved. > I personally disagree with that, and will elaborate my > argument once we've > published final minutes. Maybe the minutes should say there was no consensus on whether there was a consensus? I guess this boils down to how to interpret a consensus on direction. Some people may have been cool to the idea of sacrificing UNIX ACL models and I can understand that but nobody said we should go in a different direction (e.g. take POSIX ACL's as the base and only include NT features that fit well into that model). The consensus was on the direction. There certainly was no consensus that that would result in something that everyone could accept. That will probably depend of the details when there is a proposal actually on the table.
This archive was generated by hypermail 2.1.2 : 03/04/05-01:47:24 AM Z CST